Ampace JP30 is a new tabless 18650, 36A CDR spec, sourced from the “B grade’s” on the Chinese second hand market, seller claims the IR is about 3 mohm.
Eohuok C6 is a Sony VTC6-like battery with the same CDR(30A) and capacity specs, seller claims the IR is about 12 mohm, I think it can be seen as VTC6.
The extremely low IR does have some advantages, the voltage sag is mitigated a bit. And quite surprisingly, the extremely low IR can increase the output even in a buck driver., Since the Vf of SFT25 at 8A is very high, ordinary batteries cannot provide such a high Vf even when fully charged, but JP30 can mitigated this, so we can see the turbo output is increased even in a buck driver light.
Btw, the illumination test environment in the pictures is not ideal, the numbers themselves don’t matter, the relative changes matter.
These are very exciting developments! Although you might not have the equipment, I am curious to see what the voltage plot is like under load, and how the sag compares to other high CDR cells like Samsung 30Q.
I suspect this is just a feature of your particular combination, 8A buck with SFT25R, which has a very high Vf at 8A and essentially skips the buck phase and goes into direct drive on a full battery, so minimizing voltage sag is crucially important for maximum output. I recently swapped a 8A buck driver into a C8 with SFT25R, and noticed that the turbo output drops almost immediately (as a function of cell voltage) starting from a full, fresh Vapcell F38. With an older cell (30Q, c. 2018) the same maximum output is not attained at all even if the cell is full.
You are right, I tried it with another buck driver light(Fenix pd32v2) and Convoy’s 20% mode, and indeed there was no change. The Vf of SFT25 at 8A is too high, and the voltage sag at the full will also affect the output.
Thank you for conducting the test again to confirm!
I’ve set the 8A driver to the 1/10/35/100 mode group, where the 35/100 modes are very close to each other due to the high Vf, so differences in turbo output was very noticeable. On a good day with the F38, the 100 mode can be 1.9x brighter than the 35 mode. The 30Q from 2018 yields an increase of only 1.5x. The two modes coincide in output at around 3.6V rest voltage.
If these plots are true, that would be crazy–the discharge profile of the Ampace at 25A essentially matches that of the 30Q at 5A! It would render obsolete literally every single high CDR cell we already know.
I found some sellers on Xianyu who looked legit, not sure this thing will stay this low forever. Also, when flying in China, you need a capacity mark on the battery or it will be considered infinity Wh and denied boarding. The Chinese market actually sells these battery wraps, but ultra-low capacity ones like these for obvious purposes aren’t so easy to buy.
And btw, considering the labor costs in different places, Chinese batteries may be a better comparison, EVE 30P(30A CDR) is about 1-1.6USD.
I’ve encountered this with my 30Q’s and it’s super annoying (considering that fake Ultrafire-type cells are ok despite being much more dangerous). Fortunately there were courier services within the airport, and I got to mail them home instead of discarding them.
BTW, I’m still hoping for a throwy enough 6V30W emitter, or that some manufacturer makes a cascaded boost-buck driver. Because of the Vf issue, getting a better battery always seems like a workaround.
Yeah it’s an awkward situation for sure. I imagine another solution would be to use this high Vf throwy emitter in a larger light like L21B, whose larger battery experiences less voltage sag at the same current, so a greater proportion of its energy contributes to regulated maximum output.
I tried to keep the settings of both ceiling bounces the same, not sure why, the improvement by JP30 seemed to be greater than that in the candela test.
I also made the curves of estimated ANSI beam distance by calculating “sqrt(4*lux*a_coefficient(42))” based on the lux curves. This has little to do with the battery itself, but is mainly to show how the ANSI beam distance changes over time in a buck drive.
It seems that it is important to have tabless batteries for high Vf emitters like SFT25.Only JP30 can maintain regulation for a while, while the other batteries will quickly enter direct drive stage. The current of PD32V2 turbo is about 6A btw.
Interesting results, and a bit surprising. At 6A the SFT25R has a Vf of 3.60V, so all but one of these cells sag by at least 0.6V under just 6A of load! But HKJ’s test of the Sanyo GA cell shows a voltage sag of less than 0.4V under 7A. It’s unclear to me where this discrepancy is coming from, maybe your cells are (i) not completely full, (ii) a bit old, or (iii) maybe there are losses in the electrical path in the light.
I saw your post on reddit with additional info like submersion under water to reduce the effect of thermal stepdown. I’m sorry about what you experienced there; sometimes it can be a very difficult place for discussions that require any amount of thinking.
Do you have an estimate of what the voltage drop, say at 5A, (or total resistance) is in the purely connective parts of the circuit (i.e., springs, switches, PCBs, battery contacts, etc)? It might depend on the flashlight model, so let’s say a tailswitch S2+/C8 with Convoy’s own high current springs.
My GA may indeed be a bit old, and I’m not surprised that the various resistors and circuit logic in the flashlight make the output drop come early. This is definitely not a rigorous battery or emitter test, but at least it’s an apples-to-apples comparison, with the only variable being the battery(and the number of cycles).
edit: btw, I tried to find a newer GA to see if there is any change.
The reddit thing is kind of funny, the people who comment and vote on you in the first 2 hours probably determine the vibe of the whole post, and there may be some personal grudges, but anyway, that’s indeed not a place for deep discussions, I’m not too surprised.
